Factor Analysis (FA) is a technique that is often used to create new
variables that summarize all of the information that might be
available in the original variables. It is used to study relationships
that might exist among the measured variables in a data set. Similar
to PCA, it is a variable-directed technique. One basic objective of FA
is to determine whether the response variables exhibit patterns or
relationships with each other. While PCA produces an orthogonal
transformation of the variables and does not depend on an underlying
model, FA does depend on on a reasonable statistical model.
Multidimensional scaling (MDS) is a mathematical technique that allows
us to map the distances between points in a high dimensional space
into a lower dimensional space. It is most useful when we can map into
a two-dimensional space as this will help us visually confirm the
different class groupings. The basic principle is to reduce distances
between points in a two dimensional space.
In this talk, we will cover the basic mathematical foundations of
these two techniques and explore their applications in speech
research. For eg., statistical pattern recognition techniques such as
these are increasingly being used to differentiate between various
phones.
Additional items of interest: